Lentil (Lens culinaris Medikus) is a self-pollinated diploid (2n = 14) cool-season pulse crop widely cultivated in the Southwest Asia and North Africa (SWANA) region and India and it ranks fifth in global pulse production after common bean, dry pea, chickpea and cowpea (FAOSTAT, 2022). Globally, lentil occupies about 3 million hectares with a production of 5.8 million tonnes and an average productivity of 3 t ha^-1 (Choukri et al., 2022). In India, lentil contributes nearly one-third of global production and is the second most important pulse in the Indo-Gangetic Plains, valued for its high protein content (~25%) (Malik et al., 2022; Zeroual et al., 2022). In eastern India, including West Bengal, lentil is predominantly grown as a rice-fallow crop under residual soil moisture due to limited irrigation coverage. Delayed transplanting of kharif rice caused by erratic rainfall often leads to delayed lentil sowing, disrupting the synchronization between crop phenology and climatic conditions.
       
Abiotic stresses, particularly terminal heat and drought, severely constrain lentil productivity by impairing reproductive development, especially pollen viability (Prasad et al., 2019). Climate change-induced increases in temperature and moisture stress trigger transcriptional and physiological disruptions affecting crop growth and phenology (Sehgal et al., 2017). Delayed sowing exposes lentil to cumulative heat units, shortened day length, cold stress during early growth and terminal heat and drought during reproductive stages (Rather et al., 2022; Baidya et al., 2021). The optimum temperature for lentil growth ranges from 10-30°C; even a 1°C rise during reproduction can reduce yield by 11-13%, while temperatures above 30- 32°C impair flowering and pod filling (Venugopalan et al., 2021). Shortened seed-filling duration under late sowing results in reduced seed size and quality (Liu et al., 2021). Despite these concerns, limited information exists on the combined effects of terminal heat and drought across phenophases. Therefore, the present study evaluated lentil genotypes under different sowing dates to quantify impacts on phenology, physiology, yield and nutritional quality (Choukri et al., 2022).

Experimental design and soil-climate conditions
 
Field experiments were conducted during the Rabi seasons of 2016-17 and 2017-18 to evaluate five lentil genotypes contrasting in heat and terminal drought tolerance, selected based on late-sown performance (Baidya et al., 2021). The study was carried out at the District Seed Farm, BCKV, Kalyani, West Bengal (Fig 1), on well-drained Gangetic alluvial inceptisol with clay loam texture and medium fertility. A split-plot design with three replications was adopted, with three sowing dates 15^th November, 30^th November and 15^th December to impose heat and drought stress. Standard crop management under rainfed conditions was followed. Soil moisture and water potential were measured periodically (Richards, 1948), while weather parameters and soil moisture dynamics are presented in Fig 2 and 3.






 
Phenological observations
 
Crop phenophases, including days to germination, branching, flowering, podding, physiological maturity and harvest maturity, were recorded through periodic field observations following Erskine et al. (1990).
 
Dry matter production and partitioning
 
Dry matter accumulation and partitioning were recorded at early seedling, pre-flowering and physiological maturity stages. Three plants per replication were uprooted and separated into roots, shoots, leaves and reproductive parts. Plant components were oven-dried at 80°C to constant weight for dry matter estimation.
 
Growth and physiological parameters
 
Absolute growth rate (AGR), relative growth rate (RGR) and crop growth rate (CGR) were calculated following standard formulas (Fisher, 1971; Watson, 1952). Biomass allocation was assessed using leaf, stem and root weight ratios (Kvet et al., 1971). Relative leaf water content (RLWC) was estimated using the method of Perez et al. (2002) based on fresh, turgid and dry weights of leaf samples.
 
Biochemical traits
 
Stress-related biochemical parameters were analyzed in field-grown leaf samples collected at 50 and 80 DAS. Lipid peroxidation was estimated as malondialdehyde (MDA) content using the thiobarbituric acid reactive substances (TBARS) method (Heath and Packer, 1968). Free proline content was determined following Mohanty and Sridhar (1982). Leaf chlorophyll content was measured using a SPAD-502 chlorophyll meter at flowering and 50% podding stages.
 
Yield and yield attributes
 
At harvest, plant height, primary branches, pods per plant, seeds per pod, 100-seed weight, biological yield, seed yield and harvest index were recorded.
 
Stress response index
 
Stress response indices (SRI) were calculated for seed yield and key phenological traits using mean values from late and very late sowing relative to normal sowing.
 
Grain micronutrient analysis
 
Grain micronutrients (Fe, Zn, Cu, Mn) were analyzed by atomic absorption spectrophotometry after wet digestion of seed samples using diacid mixture, following Jackson (1973).
 
Pollen germination studies
 
Flowers collected at anther dehiscence were used for in vitro pollen germination assays at 18°C, 24°C and 30°C. Germination was recorded at 30-minute intervals up to 90 minutes, following Niles and Quesenberry (1992) and observed under a Zeiss Axiocam 305 microscope.
 
Statistical analysis
 
Data were analyzed using ANOVA for split-plot design (Gomez and Gomez, 1984). Mean comparisons were performed using Statistix 10 software. Pearson correlation and principal component analysis were conducted to assess trait associations and variability. Data visualization and analysis were further performed using Python.

Influence of sowing dates on phenological phases of lentil genotypes
 
Sowing date significantly affected the phenological development of lentil genotypes, highlighting the crop’s sensitivity to planting time shown in interaction effect of sowing dates and genotypes in Fig 4. Under normal sowing (15^th November), rapid germination and seedling emergence occurred within 5-6 days, whereas delayed and very late sowing prolonged early establishment probably due to lower soil and air temperatures (Venugopalan et al., 2021). WBL-77 showed the maximum delay in emergence under late sowing, likely due to temperature-induced suppression of metabolic activity. Vegetative stages, including first and second branching, occurred earlier under normal sowing but were altered under delayed conditions. Reproductive phases were highly sensitive, with 15^th December sowing inducing early flowering and podding owing to higher temperatures and shorter photoperiods. Delayed sowing shortened crop duration (103-106 days) compared with normal sowing (~111 days). Genotypic variation was evident, with L-13-123 flowering earliest, WBL-58 having a longer reproductive phase and PRECOZ showing extended maturity. Overlapping growth phases and accelerated development under stress were governed by temperature, moisture stress, growing degree days and photoperiod (Fig 5 and 6). Data presentation using tables could show these results better.






 
Effect of sowing dates on dry matter accumulation and partitioning
 
Dry matter accumulation and partitioning strongly influence crop yield and are affected by sowing time and environment (Table 1). Delayed sowing significantly reduced root dry weight at all stages and progressively decreased leaf and reproductive biomass compared with normal sowing (15^th November), mainly due to shortened growth duration, higher temperature, moisture stress and reduced photosynthesis (Akter and Islam, 2017). Genotypic variation was evident, with WBL-58 showing the highest root biomass, indicating better stress adaptation. Delayed sowing (15^th December) reduced total biomass by nearly 56%, consistent with earlier reports in chickpea, rapeseed, wheat and lentil (Venugopalan et al., 2021; Xu et al., 2022; Bera et al., 2024).


 
Effects of sowing dates on growth parameters
 
Growth indices including AGR, RGR, CGR, LWR, SWR and RWR were significantly affected by sowing dates and genotypes (Table 2). AGR and CGR increased with crop age, whereas RGR declined at later stages. Delayed sowing enhanced AGR and CGR but reduced RGR at 80 DAS, indicating stress-induced growth adjustments. Late and very late sowing markedly reduced most growth indices due to impaired photosynthate production and partitioning. Genotype WBL-58 maintained comparatively higher AGR, CGR and RGR under 15^th December sowing, contributing to better yield performance, consistent with Venugopalan et al., (2021). Delayed sowing increased RWR as a compensatory root response to moisture stress, while LWR and SWR declined, indicating reduced canopy development, corroborating genotypic trends reported by Akter et al., (2022).




 
Effects of sowing dates on stress indicators
 
SPAD value
 
SPAD values were significantly higher under optimal sowing (15 November) and declined under late and very late sowing. WBL-58 maintained the highest SPAD values at both 50 and 90 DAS under stress (Fig 7), indicating better chlorophyll retention and photosynthetic capacity. Reduced SPAD values in late-sown crops reflect chlorophyll degradation and nitrogen limitation under heat and drought stress (Atta et al., 2023).


 
Relative leaf water content (RLWC)
 
RLWC declined progressively with delayed sowing, with the lowest values recorded under 15^th December sowing. Higher RLWC at 50 DAS compared to 90 DAS indicated progressive dehydration under stress (Fig 7). Reduced RLWC under late sowing reflects diminished cell turgor and impaired growth (Nandi et al., 2023).
 
Lipid peroxidation
 
Late-sown crops exhibited significantly higher malondialdehyde (MDA) content, indicating increased oxidative damage (Fig 7). WBL-58 showed comparatively higher MDA levels but maintained yield, suggesting efficient antioxidative defence mechanisms (Zeroual et al., 2022).
 
Proline accumulation
 
Proline content increased significantly under late and very late sowing, particularly at 90 DAS (Fig 8). Genotypes WBL-58 and L-13-123 accumulated higher proline under stress, highlighting its role as an osmo-protectant and stress-adaptive metabolite (Brini and Saibi, 2023).


 
Effects of sowing dates on yield and yield attributes
 
Sowing date significantly influenced yield and yield-attributing traits in lentil, with highly significant effects of sowing dates, genotypes and their interactions. Optimal sowing on 15th November resulted in superior performance for major agronomic traits, including flowering time, pods per plant, seed weight, biological yield, seed yield and harvest index. In contrast, delayed sowing, particularly on 15^th December, caused marked reductions in these traits and overall seed yield. Delayed sowing particularly on 15^th December caused substantial reductions due to shortened growth duration, flower and pod abortion and impaired seed filling (Sen et al., 2016). Considerable genotypic variability was observed, with L-13-123 consistently recording the lowest performance, indicating poor adaptation to delayed conditions (Table 3). The stress responsive index (SRI) further confirmed significant declines in yield traits under delayed sowing (Fig 9).




 
Effect of sowing dates on seed micronutrients
 
There were no significant interactions between genotypes and sowing dates for seed micronutrient contents. However, the genotypic effect was significant with regards to Fe, Zn, Cu and Mn. But sowing date on its own was not significant effect. WBL-58 showed greater variability and relatively higher micronutrient levels compared to other genotypes (Fig 10). Relatively higher micronutrient accumulation due to genotypic effect is linked to impaired uptake, reduced transpiration and metabolic disruption during seed development (Ahmed et al., 2024). Environmental stresses such as high temperature and drought exerted comparatively minimal effects on this WBL-58 genotype.


 
Effect of high temperature on pollen germination
 
In vitro pollen germination declined significantly with increasing temperature. At 30°C, pollen germination was severely inhibited, particularly in L-13-123, while WBL-58 retained partial germination and better pollen tube integrity (Fig 11). High temperature during flowering adversely affected pollen viability, leading to reduced seed set and yield (Sita et al., 2018).


 
Principal component analysis
 
PCA revealed that the first three components explained 75.9% of total variation. PCA1 (48.2%) was dominated by flowering traits, PCA2 (16.5%) by pods per plant and copper content and PCA3 (11.2%) by seed weight and yield. Yield was positively associated with flowering traits, biological yield and iron content (Fig 12-13), indicating their critical role in determining performance under stress.

The study confirms that delayed sowing in lentil intensifies terminal heat and drought stress, markedly altering phenology, physiological traits, reproductive processes and yield and quality attributes. Shortened crop duration, impaired water relations, oxidative stress and reduced pollen viability collectively limited pod set, seed filling and micronutrient accumulation, leading to substantial yield losses. Significant genotypic variation was observed, with WBL-58 exhibiting superior physiological resilience, reproductive stability and dry matter partitioning under late sowing, while L-13-123 proved highly sensitive. These results emphasize the critical role of sowing time and stress-tolerant genotypes for sustaining lentil productivity under climate-vulnerable, late-sown conditions.

The authors are grateful to Bidhan Chandra Krishi Viswavidalaya for experimenting with the Doctoral Degree Programme of Ananya Baidya and thankful to the Department of Plant Physiology for providing experimental and technical support.
 
Authors Contribution
 
Ananya Baidya and Subhasis Mondal designed the study. Ananya Baidya conducted the field experiment and data acquisition and conducted lab analysis. Ananya Baidya, Subhasis Mondal, Rajeev Kumar, Ranjit Singh Gujjar, Santanu Banerjee, Kousik Atta and Nurnabi Meharul Alam conducted statistical analysis and prepared the graphs and figures. Ananya Baidya, Subhasis Mondal, Rajeev Kumar and Ranjit Singh Gujjar drafted and edited the manuscript. All authors contributed to the article and approved the submitted version.
 
Disclaimers
 
The views and opinions expressed in this article are solely those of the authors and do not necessarily reflect the official policy or position of the affiliated institutions.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.